The Duty of Violence

This essay argues that the deontological view of morality is connected to extreme and massive forms of violence through a kind of phenomenological necessity. In the first main section, I examine one family of such violence, which usually comes under the label of “religious violence”. I argue that it is not the religious element but the disqualification of context from the realm of justification which characterizes such violence. In the second main section, I examine the phenomenology of duty to conclude that duty, by definition, denies any normative relevance to context. In the third main section, I use this sketch of a phenomenology of duty to propose a hypothesis about the underpinnings of the connection between mass violence and duty, namely, that the notion of duty carries with it the exclusion of moderation, and places the agent before an impossible situation that can only be resolved by violence.


Introduction
This essay argues that the deontological view of morality is connected to extreme and massive forms of violence through a kind of phenomenological necessity. In the first main section, I examine one family of such violence, which usually comes under the label of "religious violence". I argue that it is not the religious element but the disqualification of context from the realm of justification which characterizes 1 3 such violence. In the second main section, I examine the phenomenology of duty to conclude that duty, by definition, denies any normative relevance to context. In the third main section, I use this sketch of a phenomenology of duty to propose a hypothesis about the underpinnings of the connection between mass violence and duty, namely, that the notion of duty carries with it the exclusion of moderation, and places the agent before an impossible situation that can only be resolved by violence.

"Religious Violence": What's in a Phrase?
The question of religious violence has ramifications in three areas of thought. The first has to do with the debate surrounding the relations between secular modernity and modern violence. In this context, the problem of religious violence has pride of place: it is used to prop up the ideology inherited from the Enlightenment, often with the implication that secularism is the solution not just to religious violence but to violence in general. This view is faced with the objection that modernity itself seems a fertile breeding ground for grand-scale violence (Rogozinski 2017(Rogozinski , 2020. The second has to do with the current attempts by state agents, individuals and nongovernmental organisations to prevent, discourage and contain the spread of violent ideologies, many of them religious. In this context, the question of polarization becomes most pressing: it looks like the antagonists are caught in a mirroring dynamic in which the radicalization of one (religious) antagonist leads to the radicalization of the other (secular) one. This undermines the tired view that only one of the parties is truly radical and, therefore, responsible for the confrontation (McCauley and Moskalensko 2011). The third is related to meta-ethics: how can moral discourses such as religions turn out to have immoral consequences-including consequences immoral from their own perspective such as mass violence? (O'Murchadha 2020).
These three debates constitute the background and the motivation for this paper. Although they are probably all inspired by a wish to reduce such violence, they all make arguments that are independent of this claim, and this applies to this paper as well. Regardless of anyone's normative commitments, the modern struggle to minimize widespread violence raises questions of essence such as the one I am discussing here: what is religious violence and why is it so resilient? When it comes to religious violence in particular, we encounter difficulties in establishing a firm, determinate and constant relation between religion and violence. Instead, we are faced with a dilemma: religion seems to have a close connection to both violence and non-violence so that it can have an essential connection with neither. I, therefore, begin by exploring the specific characteristics of religious violence. In response to the dilemma, I argue that religion is violent not because it is religious, but because it formulates duties.
"Religious violence" is a thick concept. It is both descriptive and normative and by "religious violence," we somehow aim at something worse than "violence" tout court. This is even more so in the current context in which any mention of religious violence immediately involves the question of how to respond to it. So, it seems that we miss something important about the way we mean to use the phrase if we 1 3 The Duty of Violence do not acknowledge the implication that the religious character of religious violence is meant as an aggravating circumstance. In turn this normative aspect of the phrase "religious violence" can be understood in two distinct ways; the first suggests that when religion becomes involved with violence the violence becomes worse; the other says that because the violence is religious it is to be deemed worse. So it seems that reconstructing what the phrase "religious violence" is meant to refer to involves addressing three questions: • What makes a certain act of violence "religious"? • Why would a connection with religion make the violence worse/more extreme? • How does the connection with religion make the perpetrator somehow more guilty?
In response to the first question, I shall argue that all the available accounts fail. There are too many senses of "religious" at work in the usual instances of so-called "religious violence" for there to be a sufficiently unified account of what makes violence religious.
Consequently, in response to the second question, I will argue that what makes the violence worse when it is religious is nothing to do with its "religious" aspect, but it is its generalizability and immoderation. Bearing in mind the response to the previous question, I also argue that these can be found in cases of non-religious violence too.
In response to the last question, I argue that there is nothing that makes the perpetrator of religious violence more guilty, but rather, that this implication in our current use of the phrase reveals not only a prejudiced attitude to religion (so that evil perpetrated in the name of religion is seen to be worse ipso facto) but more interestingly, that it contains a strategy to implicitly and a priori vindicate secularism from such charges of generalizable and immoderate violence.
I conclude that we should be formalists about violence: it is not the religious content of religious violence that should be our focus, but rather its form; generalizability and immoderation. I further conclude that this form is shared between religious and secular violence, and that what makes violence generalizable and immoderate is not its religious aspect in particular, but any indifference to context in general. I suggest we, therefore, replace the phrase religious violence with "contextless violence".

Problems with the Justification Account
One way to understand what is religious about religious violence is the idea that acts of violence are usually religious if their perpetrator believes that their actions are justified religiously. Call this the justification account of "religious violence". As O'Murchada has recently argued, this understanding is inappropriate because it fails to acknowledge the many other ways in which religion and violence can interact, besides justification (O'Murchadha 2020: 287). I would add that another problem with such a view is that it lumps together two problems that should be kept apart. The first is the meta-normative problem of what kind of justification is required for 1 3 violence. The second is the normative problem of what acts of violence are justified and why. One could point to religion for an answer to the second, but not clearly to the first. At best, one may deny that there is a question about the first (i.e., one might take for granted that religion has authority in such matters). One way to see the importance of keeping the meta-ethical and the ethical questions apart is to look at how recent developments in the theory of religious violence have largely been structured by the alternative between those who explain the wrongness of religious violence in terms of the wrong notion of justification and those who explain it in terms of the wrong justification. For example, some counter-radicalisation techniques focus on arguing the meta-normative point that religion doesn't have the authority to condone violence while others argue the normative point that the acts of violence religion condones are not those the radicals believe they are (the socalled "Islamic challenge" for example). The oft-cited debate between French Islamologists Olivier Roy and Gilles Kepel can also be, brought down to this disjunction, with Roy arguing that fundamentalist Muslims find in Islam a justification for the violence motivated otherwise (by psycho-sociological factors)-therefore, relying on the meta-normative view that religion has the authority to mandate violence-; and Kepel countering that motivation and justification come from the same place, i.e., Islam-implying that it is the contents of Islamic doctrine that induce violence.
So, the phrase "religious violence" can already be taken in two divergent senses. In the first, it refers to an act of violence which involves the meta-ethical view that religion has authority to mandate violence. In the second, it refers to acts of violence which involve that the ethical view that "religion mandates these acts". As the case of Kepel and Roy shows, it is recognized by all parties that some acts that help themselves only to one of these and not the other are common (think of the violent character who uses religion to justify their violence or of the peaceful character who is driven to violence by their faith). The problem, in fact, is even more acute: it seems the phrase "religious violence" not only fallaciously throws together a theory of justification and a set of justifications, but also that it is an operative phrase only insofar as it does so. Keep these two things separate and the specificity of "religious" violence falls apart: what we have are a number of acts of violence, some of which involve the religious meta-justification (the view that religion has authority to mandate violence) and others which involve the justification ("Scripture says so"). If we are to keep to the phrase religious violence in ways that place the blame on religion alone, we must also entertain the fallacious view that in the case of religious violence, and only in this case, these two claims are indistinct. This, however, is an ideological point, one that has a history illustrated by the history of the phrase "religious violence" itself. So it seems, if we focus on the question of justification, any specificity of religious violence qua religious collapses. This should suffice to disqualify the justification account.

Problems with the Substantial Account
Shifting our focus from the question of justification to the contents of doctrine hardly fares better. Many argue that "religious violence" refers not to the kinds of justification adduced for certain acts of violence, but to the contents of the justifications 1 3 The Duty of Violence provided for said violence. Such contents, so the story goes, are different because they are religious. We will find that in this case the notion is too coarse-grained. Consider the different position of the religious element in the following: • Violence motivated by the religious beliefs of the perpetrator when the victim has other or no religious beliefs (think of the Tokyo gas attacks); • Violence motivated by the religious beliefs of the perpetrator where the victim is of the same religion (think of the murder of Yitzhak Rabin by Yigal Amir); • Violence motivated by the religious beliefs of the perpetrator where the religion of the victim is of no concern (such as the 9/11 attacks); • Violence aimed at eradicating the religious beliefs of the victim (think of the initial wars between Shia and Sunnis in the eighth century); • Circumstances where there is a nonessential religious element in otherwise political violence (think of the figure of Abbott Cauchon); • Or conversely, where there (arguably) is a nonessential political element in otherwise religious violence (think of the Abbott's famous victim, Joan of Arc, or of the French religious wars of the 16th Century); • Violence where racial and religious motives are indistinguishable either because they're conflated or undecidable (think of Nazi Anti-Semitic violence and of the current convergence between Islamophobia and Anti-Arab racism in France); • Violence which aims to achieve a religious purpose to which violence itself is only contingently connected (think of the Arab or Ottoman wars of conquest and forcible evangelization in Latin America) • Acts in which the violence is essential to the religious purpose (either what we call "ritual violence," self-flagellation practices, human sacrifice, genital mutilation, or extermination attempts).
In other words, many kinds of violence are religious in the sense that they have a "religious" element, but none of them are religious through and through.
Add to that a further ambiguity, which lies in the ways the qualifier "religious" applies to the noun "violence". Linguists distinguish between different kinds of relations between the noun and the adjective. For example, the "syntagmatic" relation (where "religious violence" would mean "violence whose perpetrator is religious"), the "paradigmatic" relation (where it would be "violence which constitutes an aspect of religion"-something like ritual violence) and the "direct" relation (where "religious violence" would suggest that the acts in question are intrinsically religious) (Claudé 1981) among many others. Note in particular how such cases show that both the terms "religious" and "violence" refer to disparate phenomena, with the "violence" being "religious" in senses here direct and there indirect and their being religious involving here religious affiliation, there religious beliefs, here what Derrida calls "the religion" ("the religiosity of the religious"), there what he calls "a religion" (Derrida 1978: 96), here the religion of the victim, there the religion of the perpetrator. In other words, many kinds of violence are religious in one of the ways an adjective can apply to a noun, but none of them are religious in any single one of these ways, or in all of them.

3
So, even as we hold ourselves to the uncontroversial cases of religious violence, we are left not only with an irreducible diversity of kinds of religious violence (I assume such diversity is what O'Murchadha (2020) alludes to in p. 288) but also with an irreducible diversity of ways in which the adjective "religious" applies to the noun "violence". I conclude that many cases legitimately associated to religious violence are religious in different ways, violent in different ways and sometimes not religiously violent at all.

Problems with the Pragmatic Account
So, it looks like we are about to conclude that the phrase "religious violence" is misleading and relies on a category mistake. Before we do so, we might try one last strategy. One might choose to dismiss the inconsistencies mentioned above in pragmatic terms and go the family resemblance route. Diversity of literal senses doesn't disqualify a more or less continuous practice which brings together, for all-or most-intents and purposes this motley bunch of notions into one linguistic practice, one phrase which can be significantly used. This of course is only a viable option if using the phrase "religious violence" still gets us what we pragmatically expect from it. This is a risky route to take. First, the phrase "religious violence," as I mentioned above, although it may conceivably be different in other possible contexts, is, in our current language game, a thick concept. In other words, we use it because we believe that using it plays its part in the effort to alleviate the kind of horrors that we associate with it. This alleviation may be expected to be achieved through rhetorical, legal (one may imagine laws that consider "religious violence" to be an aggravating circumstance), or theoretical means (for example if using it helps philosophers and social scientists understand it better with a view to responding to it better). Correlatedly, the reason religious violence is near the top of the list of things to address in such performative ways is related to the fact that we think of it as an especially urgent problem. This also suggests that the adjective "religious" in this context is meant to signify that this is violence somehow made "worse" by its religious element. In this context, the family resemblance route is little more than a copout: lumping together all the different senses of religious violence enumerated above, along with all the different ways in which "religious" qualifies violence, is not only like bundling apples and pears (which the family resemblance view can accommodate), it also fails in its own project: to understand and address the especially severe phenomenon we aim at behind this phrase. Indeed, what must urgently be identified are the mechanisms that make religious violence such an urgent problem. So, the family resemblance route doesn't redeem the phrase "religious violence" but it does allow us to make a step forward: "religious violence" is a thick concept because it has a pragmatic dimension. I suggest we police our use of concepts so as to account for the kind of violence that is of current urgent concern in ways that allow us to cope with it.

3
The Duty of Violence

An Alternative Proposal: "Contextless Violence"
Although I have suggested above that the intuition that singles out religious violence by way of its religious component is mistaken, it seems to me that it is correct in its implication that there we have a kind of violence that is horrifying. I think the valid insight that lies behind this intuition is the following: religious violence is worse than other kinds of violence because it is generalizable and immoderate. Although elucidating the implications of characterizing religious violence as generalizable and immoderate would take us far afield, seeing that this is the case should not take too much effort.
After all, it is quite commonly claimed that religious terrorism (for example) is "indiscriminate" (it doesn't distinguish between combatants and civilians, children, the elderly, women and men) or as an Islamist militant put it "there is no red line" in terrorist action (Speckhard 2012). Similarly, religious violence, like religion (at least monotheistic religion) itself (mutatis mutandis) has a universal vocation, that is to say, it is justified everywhere if it is justified anywhere. Finally, most acts of religious violence are motivated by doctrine which is itself regarded as authoritative through time and space and/or by tradition whose authority is increased by temporal distance. Sacred texts for example have no vocation to be adapted to the different epochs in which they are used. What underpins generalization and immoderation, therefore, is the religious indifference to context. Religious violence in one context is easily (if not automatically) transferable to other contexts (hence the current "globalization of Jihad" for example) because its justification is not related to these contexts. Contrast this with other kinds of violence: say, violence in sports (like in boxing), that is entirely dependent on the context provided by the magic circle of play, but also domestic violence, which, although extremely severe and an aggravating circumstance of its own, is most often carried out because it is triggered rather than justified. Similarly, the justification for violent robberies, defensive violence or even most wars in which the appeal to violence is purely instrumental expires with the context, where the relevant change of context is called "victory" or "defeat". The connection of religion with generalization applies both to the normative underpinnings of said violence (they are unaffected by time and place) and to their victims. We are all horrified to observe the religious fanatic's ability to treat their victims interchangeably (i.e.: their ability to regard individuals as potential victims, or enemies in principle) revealing that the violence refers to an ideal, not to a specific victim.
This leads to its second characteristic: immoderation. It seems the intensity of religious violence suggests that the amount of suffering it causes is not considered relevant to the justification of said violence: what the violence is trying to suppress or destroy is an abstract principle, one that cannot be suppressed or destroyed piecemeal. The moderating principle which would involve reluctance when the violence required is deliberately silenced. So, it seems religious violence is immoderate because it doesn't accept any contextual factor as normatively relevant, and it regards all moderating factors as contextual (moderation, as Aristotle well saw, is always a matter of in situ judgment). 1 So, it seems we have good pragmatic reasons for redirecting the attention once devoted to religious violence alone towards all instances of violence that is generalizable and immoderate. I suggest we take a "formal" approach, therefore (Bronner 2009; Chouraqui 2019), one that isn't staked on the contents (religious or not) involved in any specific violent act but rather focusses on the imperatives of generalizability and immoderation. Following on my earlier work (Chouraqui 2021), I suggest that generalization and immoderation should be understood in a unified manner as an indifference to context. Consequently, the kind of violence we need to stay focused on without being distracted by the question of religion should be formally defined as characterized by its indifference to context. Now, although contextlessness is arguably related to some basic structures of the religious mindset, they are not specific to it. On the contrary, if I am right to claim that contextlessness is the crucial point here, this means that the seed of violence is lodged at a level that precedes the disjunction of religion and secularism. In this context, we should be critical of the current doxa in social theory which makes the genealogical argument that secular power has a religious provenance or which invokes "the sacred" as the quality that is genealogically transferred from a religious context to a secular one. There is, I argue, no genealogical precedence of religion over secular contextless violence. 2 What we need, rather, is to find the common source that such contextless violence in its secular and religious forms, share. 3 This also means that the focus of our investigation must be transcendental (and, therefore, philosophical): we must ask ourselves a) what the a-historical source of such tendency towards generalization and immoderation come from and b) how this tendency towards these two abstract notions results in actual violence. I address a) in the following section and b) in the section "Duty and Mass Violence are Necessarily Connected" further below.

3
The Duty of Violence

Phenomenology of Duty
Talking about "contextless violence" represents progress over the initial fallacious notion of religious violence, but it remains a misnomer itself: it cannot be that violence is without context since violence is always in situ. So, we must ask what exactly is contextless if not the violence? The foregoing seems to make clear that what makes contextless violence contextless is the claim that accompanies certain acts of violence to the effect that the context is normatively irrelevant. So, contextlessness refers to a claim about normative relevance that accompanies certain actions. In short, it is a claim that says that the violent action is appropriate regardless of context. Appropriateness in turn refers to justification and obligation: justification and obligation are taken to be indifferent to context.
Contextless violence exists because the claim that context is normatively irrelevant is somehow viable enough to underpin certain violent actions. In order to analyse the mechanism of contextless violence, therefore, we must pay attention to the kinds of discourses that present context as normatively irrelevant to the justification of violent actions and to the conditions according to which they indeed succeed in motivating such actions. This means that we must isolate the implicit beliefs illustrated by contextless violence in terms of the motivating experience that they induce in the perpetrator of such violence.
In our conceptual tradition (the tradition of Western Enlightenment rationalism), the chief example of a transcendental justification for dismissing context as normatively irrelevant lies in Kant's notion of duty. Duty, Kant famously argues, is "unconditional," a term used coincidentally by contemporary scholars when they analyse extremism (e.g., Bronner 2009), because it is transcendentally warranted. In the terms of the preceding discussion, this suggests that we must investigate the experience of duty that perpetrators of contextless violence undergo. The rest of this section is devoted to a phenomenology of the pathos of duty.
Perhaps surprisingly, I am not aware of any explicit phenomenology of duty, although some starting reflections may be found in Scheler (1973) and Wojtyla (see Modras 1980). Implicitly, but more importantly, we have Kant's own musings in the Groundworks, Theory and Practice, the Second Critique and the Metaphysics of Morals. I will begin by attempting a rough sketch of a phenomenology of duty based on these Kantian remarks. I will focus on one implication of Kant's account, namely that the pathos of duty entertains an ambivalent relationship to interpretation because of its ambivalence about context. For Kant, the problem is the following: duty always involves the duty to ignore context, and yet the demands of duty only arise in context. I will then argue that this tension underpins the affinities between duty and contextless violence and, therefore, that their relations are not merely accidental. I conclude that the imperative of violence reduction is suited not to a rationalist morality of duty but to a hermeneutic ethics. In this section, I will restrict myself to three basic points. Firstly, for Kant, the experience of duty involves that empirical reality is not normatively relevant. Secondly, and by contrast, for Kant, the experience of duty involves a commitment to the world, and thirdly, for Kant (as is well known), the experience of duty involves a new experience of freedom.

Three Commitments to Reality
The most obvious characteristic of deontological ethics is that it considers empirical circumstances (either context or consequences) as normatively irrelevant: the value of an action doesn't depend on them. In fact, duty, Kant argues, is indifferent to even personal well-being (Theory and Practice, 282, passim), and even mandates that one abstract from concrete circumstances (Theory and Practice,283). This is what Nietzsche calls the "cruelty" of the Categorical Imperative, (Nietzsche 1999, section II, 6). However, as Hegel (1986, Preface, Sect. 5, appendix, see also Hoy 1989 already pointed out, this only leads the Kantian into an ambivalent relationship with the concrete dimension of practical life, because even as it denies the relevance of in situ circumstances, Kant's deontological ethics cannot do away with a commitment to concrete and contingent practical life (Russon 2004). In Kant, this commitment takes three forms: situationism, objectivism and personalism. I will briefly sketch out what these three commitments entail to conclude that this leaves Kant in an ambivalent position, which he resolves by committing to contextlessness.

Situationism
Kant holds that duty has a situationist aspect, that is to say that a) although duty ignores specific contexts, it aims at them, and b) although the moral demand subsists outside of any context, it only arises on the basis of a given situation. As John Russon puts it: "Moral action […] only exists when it has an immoral, irrational other [the contingent situation] to oppose." (Russon 2004: 148), in other words, for the moral law to demand a moral action is to demand for it to be instantiated in the world. As Kant himself puts it: "it would not be a duty to aim at a certain effect of our will if this effect were not also possible in experience" (Theory and Practice: 280). This is why, of course, duty is always the duty to have a certain comportment or carry out a certain action. Doing an action for the sake of the law is always doing this action, afforded by the world and informed by the world. For all his emphasis on the indifference to context and consequences, Kant remains committed to the idea that, without instantiation, morality is idle and that an idle morality is morally insufficient. In other words, any theory of duty will have to assume that it is, in every case, for me to carry out this duty, and this is the case because I am in the relevant situation in ways that allow me some access to it, and that the Law left to its own device, isn't (Critique of Practical Reason: I. 20. 1). Conversely, this suggests that the demand for morality is always double: the demand to respect the law involves "do your duty" and "do action X informed by the circumstances". Let's take the standard example of the duty to tell the truth. If the revolutionary guard demands that I inform on my friend who is hiding in my basement, my duty, to tell the truth, becomes cashed out into the duty to say: "my friend is in the basement." This is, of course, informed by the empirical situation: it is so only because my friend is in the basement, and because I have been asked.
As a result, respect for the law is not complete without the establishment of a seamless and organic unity between the transcendental and the empirical realm, between our commitment to do our duty in general and our commitment to carry out The Duty of Violence this action in particular. This duty to "cash out" the general precept "do not lie" into the specific situated act of informing the revolutionary guard is left to the subject's discretion, although Kant covers this up with talk of "sakes" and "maxims," we'll return to this.

Objectivism
The second way in which our experience of duty relies on our commitment to reality concerns Kant's objectivist account of the Law. It is famously unclear how this objectivism is to be understood, and constructivist and realist readers of Kant have long discussed whether the objectivity of the law makes it a noumenon or merely a transcendental necessity (Ameriks 2003;Bagnoli 2015;Formosa 2013;Hills 2008;Kleingeld 2010;Rawls 1980;Stern 2007;Stieb 2006;Tiffany 2006;Westphal 2003Westphal , 2016. For example, one may wish to debate whether a human consciousness must believe in the noumenal character of a law in order to obey it, or if Kant can rely on some more parsimonious motivation (as he seems to do in Critique of Practical Reason, 1.4 in particular). I don't think I need to take any position in this debate in order to make my point which is that, however "objective" and "exist" are to be understood, the law at least exists as bindingness. As is well-known, the experience of the law is most often called "awe" and "reverence" (Theory and Practice 285, Critique of Practical Reason 5: 42f., 19.1). Another famous concept, "respect," is a synthetic experience that contains three elements 1) the recognition of the fact of the Law, 2) the recognition that the Law exists at least in the mode of bindingness (as a "fact of reason" Critique of Practical Reason: 19.1), and 3) it inspires in us the motivation to fulfil, honour or obey, the Law. In short, phenomenologically speaking, the theoretical recognition that there is the law, and the practical obligation to carry it out are one thing, unified under the term "respect". Obedience, therefore, should not be construed as following from respect, but it is identical to it. More importantly for our purposes, even a charitable account of Kant's objectivism (such as the constructivist one), will commit Kant to describing the bindingness of duty as requiring the existence of the Law. The debate is not there, but on the question of what mode of existence should be assigned to this Law.
Wherever one stands on the debate between realists and constructivists, that is, on the ontological status of the Law, it is clear that the question of the grounds for its bindingness is left underdetermined: there is no objective fact to support it. For the realists cannot claim that the Law is binding by virtue of its existing noumenally (this would immediately lead into the is/ought fallacy) and the constructivists cannot claim that the Law is binding just because it is grounded transcendentally (this would make it impossible to distinguish it from an idea of reason, and this would commit them to confusing theoretical necessity with practical necessity). Bindingness, in other words, cannot depend on either the state of the world or the state of the subject (Longuenesse 2012: 32). This pushes Kant into some sketchy but deep remarks about the simultaneity of theoretical and practical necessity, neither of which precedes or follows the other as they are unified in the one first truth: the transcendental unity of apperception. In other words, it leads to Kant's commitment to personalism.

Personalism
For Kant, therefore, my commitment to the moral law coincides with the character of my being at the most fundamental (transcendental) level. In general, Kant describes this character as the transcendental unity of apperception, but in his moral philosophy, this "invisible self" (Critique of Practical Reason, Conclusion 1.6) is further fleshed out as a commitment to unity in such a way that the transcendental unity of apperception can now serve as a ground for both theoretical and practical reason. It is this unity, understood as both a fact and a demand, that underpins both our personal identity and our commitment to the law at once. This demand, Kant calls "personality" and he equates it explicitly with duty, for "duty is nothing else than personality" (Critique of Practical Reason:, 22.2;Longuenesse, 2012: 20;Seung 2007) and it is the object of respect (Critique of Practical Reason: 24.1). This should address the question of bindingness: the law is binding not because it exists noumenally or because we are constitutively committed to it (see Longuenesse 2012: 32), but because our commitment to it is immediately implied in the transcendental unity of apperception (see also 1.4 in 2nd Critique 34). 4 From a phenomenological point of view, this is crucial because it confirms that the feeling of duty is not secondary to any experience. Rather, as a fundamental aspect of the transcendental ego, it is a condition of all experience and, therefore, a structure of the world itself. Its phenomenological primacy is what grounds its normative authority. As a result, one must admit that duty is only binding if it is grounded in the unity of recognition and institution: it is the transcendental ego, not the empirical ego, that underpins justification. And this means that justification cannot do away with arbitrariness: duty is dependent on a theoretically (not practically) necessary fact, the transcendental ego. Longuenesse herself calls this Kantian move a "petitio" (Longuenesse 2012: 24).
The pathos of duty which animates Kant's moral theory involves the simultaneous commitment to objectivism, situationism and personalism. As a result, this is a pathos that entertains an ambivalent relation to the real world. To speak more phenomenologically, Kant's view exhibits a certain resource in human consciousness, namely that one can at once consider the transcendental world as the standard for the empirical world (because the contingent world is not as real as the Law), and the empirical world as the standard for the transcendental world. Everything takes place as if the Law is not real enough until it becomes instantiated in the contingent order.

3
The Duty of Violence Crucially, it suggests also that this unity of opposites is not the kind of lived contradiction that only poses problems for the logician while it is constantly resolved in practice. For the resolution in practice itself presents a difficulty: every time I arbitrate between the demands of the contingent situation and of duty, I at once resolve the tension at the heart of the pathos of duty and violate the demands of duty (for it demands I ignore the contingent world). In short, it is a tension that can only resolve itself via individual and unlawful interpretive choices.

Freedom as Autonomy
This leads into the second basic phenomenological structure of duty, which is a new notion of freedom. Famously, Kant inaugurates the notion of freedom as autonomy or obedience to the Law. The last two points, on situationism and on personalism, are two sides of the same coin because respecting the law and maintaining one's unity are one existential commitment. 5 More generally, this means that carrying out one's duty is the only way to be oneself, or rather, that it coincides with being oneself. In Kant's language, it means that we are autonomous: only what we are dictates what we do. Kant doesn't tire of dramatizing the paradoxical form of this new notion of freedom: Duty, he argues is, "the maxim of unconditional observance of a categorically commanding law of free choice" (284) and famously (21.1).
Duty! Thou sublimely great name, that encompasses nothing endearing which entails flattery, but rather demands subjugation, but still also threatens nothing which arouses and terrorizes natural abhorrence in the mind in order to move the will, but rather erects merely a law, which finds entrance into the mind of itself and still acquires reverence even against the will (though not always compliance). (Critique of Practical Reason: 21.1) As Findler puts it, what Kant has in mind is an experience of freedom which involves "I am both my own master and slave as an autonomous being" (Findler 1997: 14). But, as Freud (via Longuenesse 2012: 31) shows, in the context of duty this freedom involves alienation for it relies on our universalising ourselves, our distantiating ourselves from our particular dimension. Of course, Kant believes that the paradox is only apparent: that it is only to those who consider the empirical side of our person as an integral part of our "personality" that the subjugation of the former for the sake of the latter will appear as paradoxical.
However, beyond the paradox of merging freedom and obedience lies a more important sticking point: Kant argues that the notion of freedom as obedience is the only available one: no act is free unless it is an act of obedience (to the Law). This contains a phenomenological commitment to a certain view of the essence of the world: namely, that one can live in the world freely without ever having to carry out an action that is not predetermined by the preexisting, permanent Law. In short, this presupposes that the empirical, contingent realm, is either the subject of a law (pace the famous Aristotelian dictum), or that there is no such contingent realm in the relevant sense (which amounts to the same).
Yet, on the basis of other parts of Kant's own works, neither consequence is acceptable: the dismissal of the contingent world is quite patently inconsistent with the reliance on the transcendental-empirical constitution of the subject who is charged with instantiation (see above section Situationism), and the notion that there can be a law of the particular is explicitly dismissed in Kant's argument for formalism in ethics.
Therefore, one can say that Kant's basic phenomenological sketch of the pathos of duty is characterized by its ambivalent relation to realism. This ambivalence introduces a tension between the duty to act for the sake of a situation and the duty to abstract from the situation. It also introduces a tension between the transcendental and empirical sources of normativity. This double ambivalence is meant to be resolved by a moment of interpretation: enacting the law in the real world is meant to be achieved by way of the identification of the normative stakes of the current situation and of the maxim of the action called for and implemented. Yet, this moment of interpretation is meant to be, itself, subjected to a universal law; anything less would involve taking the law in our own hands and violating autonomy. These tensions culminate in the tension between the obligation and the interdiction of interpretation. Even when all goes well, honouring duty requires input from the situated, empirical individual, and it is only the situated individual who can determine appropriately-and, therefore, without rule-what the situation means morally, and what the moral law requires situationally (Kant comes close to such an admission in the Critique of Practical Reason: 13.1).
Therefore, we can say that the notion of duty counts on interpretation (as the ability to abstract maxims from acts), to determine the appropriate course of action in a given situation, and to read the situation against the demands of the moral law. But it also rejects it as giving away too much to context, and as inviting the wrong kind of freedom. This is what "carrying out an action for the sake of duty" or what extracting the maxim of an action amounts to. For extracting a maxim from an action or an action from a maxim takes an act of interpretation which itself-just like any act of interpretation-cannot be regulated by any categorical imperative. The covering up Kant and any deontological doctrine will have to commit lies in overlooking this spontaneous, lawless and subjective moment by assuming that maxims are self-evident and, therefore, that carrying out one's duty only implies obedience, not interpretive choice (see note 125 of the Second Critique).
This leads us to conclude that the pathos of duty corresponds rigorously to the pathos of contextless violence as characterized in section one. That is to say that it presupposes the existence of a self-evident Law whose rules of applications are universal and self-evident themselves. As a consequence, the affinity between duty and contextlessness can be made more precise: duty always involves the duty of decontextualization alongside the interdiction of personal decision. But the phenomenological analysis of duty allows us to take this yet one step further: with the duty of decontextualization lies the interdiction of interpretation. This is best expressed by paying attention to the notion of freedom as autonomy which is requisite for any deontology. For the interdiction of taking the context into account involves the interdiction of interpretation (the taking into account of context is by essence interpretation) and interpretation engages a kind of freedom irreducible to autonomy: hermeneutic freedom. Contextless violence as described above and the pathos of duty, therefore, share the same essence: the covering over of hermeneutic freedom. In the next and final section, I investigate the role of this moment of repression of hermeneutic freedom in generating violence.

Duty and Mass Violence are Necessarily Connected
Now that the affinities between duty and contextless violence have been established, I shall use this section to examine exactly the status of these connections. I will argue that they are bound by both a causal-psychological and a phenomenological necessity. In particular, I argue that the ambivalent demands of the pathos of duty which at once demands and bans the exercise of hermeneutic freedom, sub-tends the phenomenological bond between mass violence and duty. Although they are not, as a Kantian would fear, a priori necessary, I argue that phenomenological necessity of the kind I observe here is all that the charge of complicity with mass violence needs.

Causal Connection
Before returning to questions of essence, it seems to me that some causal connections between contextless violence and the pathos of duty can be extracted at the level of the first characteristic of the phenomenology of duty, that is to say, the experience of ambivalence towards reality. It seems that this would be the kind of connection that Hegel (Rechtsfilosofie, Preface, Sect. 5, Appendix;Robinson: 1977) and Freud or Ricoeur for example, would point out. Under the heading of "pathologies of duty," they would argue that ambivalence in general provokes violence (Ricoeur 1970: 164, see also Wahl 2008. The argument, as is well known, goes thus: if the world is both the standard of value and the object upon which a standard must be applied, then it is forbidden to judge it and not to judge it. This places the moral subject in an impossible position and in effect, in competition and ultimately in a struggle to the death with, the world. As Merleau-Ponty points out in a text which begins with the phrase "Hegel says that the Terror is Kant put into practice," this leaves us with the alternative of destroying ourselves or destroying the world. (Merleau-Ponty 1948/1990. The experience of duty is central in this structure because it induces the belief that it is impossible to act morally and impossible not to act at all (indeed, doing our duty requires it). That is to say that duty tells me that I am guilty for the world being such that I need to break the Law in order to enact it. This takes resentment to the ontological level, which is Hegel's concern: I become a hater of the world. This conflict creates resentment and a zone of lawlessness: the agent finds themselves under the in situ necessity to act, and, therefore, pick a side, leading to guilt and, therefore, the need to engage in radical evil: to rationalize what we did by remaining faithful to it (similarly, William James (2012) defines fanaticism as "loyalty taken to the extreme"). Our psychological makeup is such that before such contradictions, our response can only be one of constant guilt (for interpreting or for not interpreting respectively), disempowerment and alienation (Longuenesse 2012: 31). As noted above, we tend to respond violently to ambivalence.
It seems to me that this account, although correct, remains too tainted by psychology to offer a robust rebuttal to those defenders of the morality of duty, most of whom would be prompt to point out that such psychological contingencies (whereby conceptual ambivalence induces violent affects) cannot invalidate the transcendental necessity of obeying the Law.
In fact, Kant himself preempts such objections with his notion of "radical evil." (Kant 1996). Certainly, so the Kantian story goes, a sense of duty appended to a wicked maxim will make the resulting action worse, but this can only happen when the maxim is misconstrued (misinterpreted), and in turn, this only happens when we take our inclinations to be our duties (see also Seung 2007). If this response is to count as a defence against the charge that deontology induces extreme violence, this would have to suggest that rigorously speaking, the relation of duty and violence is one of mutual exclusion rather than one of collusion. Can shifting the blame to irrationality bear the brunt that Kant needs it to in order to clear duty of its association with violence?
The radical evil argument cannot suffice however. It is itself circular because it presupposes rather than demonstrates that the evil that comes from the pathos of violence is due not to deontological demands but to inclinations. In order to do so, it must define inclinations as the source of evil, rather than discover that all evil comes from them. More importantly and leaving aside circularity, the argument is weak. First of all, Kant himself insists that the categorical imperative informs what we should call moral, and doesn't conform to what we, for contingent reasons, do call moral (including for example violence-avoidance). In fact, Kant is content to admit that deontological ethics can have counter-intuitive results. The example of lying is striking in this connection, for Kant concludes that one must give one's friend away to an unjust authority in order to avoid lying. Secondly, Kant insists that the categorical imperative offers a formal criterion for moral action. The categorical imperative is first and foremost a criterion of consistency. By extension, it is a criterion of rationality, but not a criterion of non-violence. In other words, the categorical imperative cannot be of the kind: "do not commit violence". If violence-avoidance is to conform to the categorical imperative, it will be due to a mere coincidence. In fact, such a commandment of violence-avoidance isn't even implicitly required by the categorical imperative, for all depends on the maxim of our action and this, as I've suggested above, is confusing and underdetermined to say the least. Additionally, Kant himself seems to make room for violence not only in his political texts but in his moral texts too. There is no reason therefore to assume that, even coupled with the categorical imperative, the pathos of duty will always comply to the imperatives of violence-reduction.
This very quick examination of the strategic use of the notion of radical evil, therefore, shows that a) sticking to the transcendental level in order to assess the appropriateness of deontological ethics is question-begging and in fact untenable even from the point of view of the deontologist herself and b) that nonetheless, a purely psychological argument will not be sufficient. In order to be of any help, the connection between duty and contextless violence must not lay content with psychologizing accounts. Between these two poles, the transcendental and the psychological, lies the phenomenological. I suggest that we have good reason to look for the connection between extreme violence and duty at the phenomenological level. If so, this would suggest that their connection is necessary enough to justify blaming deontology for extreme violence, but not so much that it can be done so a priori. In short, the connection between violence and duty lies in the implementation of duty, which is, as you will recall, a transcendental necessity of duty itself (see above section Situationism). As a result, and although the connection is not itself transcendental, it obtains for all possible applications of duty. This should give us an amount of necessity sufficient to make us evade the charge of psychologism.

Phenomenological Connection
The easiest way to establish such a phenomenological necessity is negative. In this last step, I will attempt to demonstrate not the connection between contextless violence and duty, but rather the necessity of the conflict between the imperatives of duty and the imperatives of moderation. The phenomenology of duty involves a reliance on a hermeneutic faculty at two important junctures: maxim-determination (for application) and situation-appraisal (for the arousal of action). It also relies on the denial of this double moment of interpretation which it covers up with an appeal to self-transparency: maxims are presumed self-evident and the moral status of a given situation too. 6 In short, duty relies on the notion that the normative meaning of an action (its value) is transparent, objective, determinate and intrinsic to the action. In this sense, the Kantian sees no room for interpretation: to be an agent is simply to pick up on the intrinsic properties of the proposed action or of the current situation. The metaphysical presupposition of deontology is, therefore, that the demands of duty are in every case determinate and transparent. For deontology, the only possible notion of freedom is autonomy, and recall, autonomy is defined by obedience to a preexisting Law. In other words, the moral world is always determinate and every moral entity (action, situation, will etc.) objectively contains the determinations sufficient to allow freedom as obedience to produce the appropriate action: there is no morally equivocal situation, the very notion of freedom as autonomy depends on this. This is obviously unsound from a hermeneutical perspective. Any of the main theories of interpretation show that the very idea of objective meaning is contradictory and naïve (on the implications of this view in the normative context and how it conflicts with Kant's metaphysical presuppositions, see Hirsch 1982). Rather, meaning-assignment only takes place where the objective state of affairs leaves at least two interpretations open, requiring an active input on the part of the interpreter. For a Kantian, this input can only be unlawful: it makes use of an unacceptable kind of freedom. In consequence, the hermeneutic approach always retains a certain notion of freedom as arbitrary. This means that deontology presupposes that the meaning of facts is self-evident and complete so that the acts of interpretation can be construed as determined responses to them, not active and arbitrary choices. In short, any appeal to deontology presupposes a univocal ontology and hermeneutics involves an equivocal ontology. The deontological view leaves the subject before the interdiction to determine the meaning of a given situation that demands action or the meaning of the action they propose to undertake. Because of its inescapable commitment to the notion of freedom as autonomy, deontology excludes hermeneutic freedom. 7 Yet, the duty to act forces the deontologist towards a decision which they are unable to fully justify. As I argued above, any careful look at the hermeneutic steps involved in any implementation of duty will have to recognize that such a complementing of autonomy with arbitrary freedom is not only unavoidable but required. This leaves the subject before a morality that both requires and prohibits interpretation all at once.
There lies the phenomenological necessity of the relations between violence and duty. The transcendental demand for empirical implementation means that our duty is to violate duty, resulting in conflict and dilemma for the moral subject. This is only a consequence of the Kantian ambivalence towards realism: because we must act in situ, interpretation is necessary; because duty is indifferent to context, one can regard autonomy as the only normative guide. Agents are now faced with an alternative: their first option is to make an existential decision in favour of implementation, thereby stepping into where no moral principle applies (recall that deontology admits of no other guidance than the guidance of duty), and, therefore, to naked interest. The situated subject, abandoned by their moral principles, is left with a world where the Law demands absolute obedience but remains silent, leaving her constantly grasping for balance. The second alternative is rather to respect the taboo of interpretation, leading to a rejection of the empirical world and thereby relinquishing any kind of action at all. The result in both cases is, paradoxically, a silencing of morals, either through unguided action or moralistic passivity (Merleau-Ponty 1948). It is important to emphasise that unlike the psychological response described in III. 1, this is a predicament that arises from the very nature of moral action, which is to instantiate general principles in experience. The process through which deontological morality produces violence is, therefore, not merely a causal and contingent one: by separating sharply the demands of the situation from the transcendental demands of the Law, and by granting absolute priority to the latter, the pathos of duty de facto disqualifies moderation. Moderation, as a long Aristotelian tradition suggests, is only a situational modifier for an otherwise non-contextual command 1 3 The Duty of Violence (Chouraqui 2021): it is always in situ, cannot be reduced to a law, and, therefore, it depends on what I have called above "hermeneutic freedom": establishing whether the current situation warrants moderation (and how much moderation) can only be left to the hermeneutic subject. By repudiating hermeneutic freedom, the deontological view repudiates any possibility of moderation.

Conclusion
Violence is religious to the extent that religion promulgates duties. Therefore, a careful analysis of religious violence will focus on the affinities between the phenomenon of duty and the phenomenon of violence. Yet, not all duties are founded religiously. The gravity of the resulting violence, therefore, cannot be uncritically traced back to its religious dimension, but rather, to its formal commitment to immoderation and generalizability, that is to say, to contextlessness. In duty, both religious and secular, context must be ignored so that the appropriateness of violence is indifferent to circumstances, and that is the source of generalizable and immoderate violence.
A closer analysis of contextlessness shows that the horrors of mass violence are inextricably linked to a commitment to deontology. More precisely, the problem lies with one specific aspect of the pathos of duty: its ambivalence towards interpretation understood as the faculty for contextualization. As Kant's implicit phenomenology of duty suggests, both the ban and the demand for a moment of interpretation are inalienable parts of deontological ethics. Mass violence and the pathos of duty constitute, therefore, an irreducible organic unity which is soldered via hermeneutic naivety (the belief in self-evident interpretation). This provides insights into the structure of contextless violence: it is violence that seeks self-legitimation through the fantasy of direct interpretation and objectivism (including universal approval), seeks redemption from the sin of interpretation (guilt), and, therefore, seeks restoration of the integrity of the unified self.
This may suggest a few final additional insights to address the ideological background of the current discussion: • First, it seems that the dominant position of deontological ethics in modern postenlightenment ideologies might now be cast not as the response to obscurantist mass violence, but as complicit with it. In particular, if this is true, this might offer a few interesting avenues of research regarding the dynamics of polarization between liberal democracies and religious fanaticism. There, the suggestion would be that this polarization must come from a shared assumption in the opposed parties. • Second, the current ideological context in which western-style Enlightenment ideology claims more or less implicitly to be a rampart against and immune to obscurantist violence has been established on the back of at least two fallacies: the first is the one contained in the phrase "religious violence" to denote "contextless violence" and the second is the "radical evil" argument which offers an unfalsifiable objection to any critique of deontology. If mass violence is to be addressed, deontology will have to be questioned, alongside the cover-ups that have hitherto enabled it to escape criticism.
Funding None.

Conflict of interest
The author declare that there is no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.